Computational instantaneous wave-free ratio (IFR) for patient-specific coronary artery stenoses using 1D network models

Computational framework for the generation of one-dimensional vascular models accounting for uncertainty in networks extracted from medical images

One-dimensional (1D) cardiovascular models offer a non-invasive method to answer medical questions, including predictions of wave-reflection, shear stress, functional flow reserve, vascular resistance and compliance. This model type can predict patient-specific outcomes by solving 1D fluid dynamics equations in geometric networks extracted from medical images. However, the inherent uncertainty in in vivo imaging introduces variability in network size and vessel dimensions, affecting haemodynamic predictions. Understanding the influence of variation in image-derived properties is essential to assess the fidelity of model predictions. Numerous programs exist to render three-dimensional surfaces and construct vessel centrelines. Still, there is no exact way to generate vascular trees from the centrelines while accounting for uncertainty in data. This study introduces an innovative framework employing statistical change point analysis to generate labelled trees that encode vessel dimensions and their associated uncertainty from medical images. To test this framework, we explore the impact of uncertainty in 1D haemodynamic predictions in a systemic and pulmonary arterial network. Simulations explore haemodynamic variations resulting from changes in vessel dimensions and segmentation; the latter is achieved by analysing multiple segmentations of the same images. Results demonstrate the importance of accurately defining vessel radii and lengths when generating high-fidelity patient-specific haemodynamics models. KEY POINTS: This study introduces novel algorithms for generating labelled directed trees from medical images, focusing on accurate junction node placement and radius extraction using change points to provide haemodynamic predictions with uncertainty within expected measurement error. Geometric features, such as vessel dimension (length and radius) and network size, significantly impact pressure and flow predictions in both pulmonary and aortic arterial networks. Standardizing networks to a consistent number of vessels is crucial for meaningful comparisons and decreases haemodynamic uncertainty. Change points are valuable to understanding structural transitions in vascular data, providing an automated and efficient way to detect shifts in vessel characteristics and ensure reliable extraction of representative vessel radii.

Computational framework for the generation of one-dimensional vascular models accounting for uncertainty in networks extracted from medical images

One-dimensional (1D) cardiovascular models offer a non-invasive method to answer medical questions, including predictions of wave-reflection, shear stress, functional flow reserve, vascular resistance, and compliance. This model type can predict patient-specific outcomes by solving 1D fluid dynamics equations in geometric networks extracted from medical images. However, the inherent uncertainty in in-vivo imaging introduces variability in network size and vessel dimensions, affecting hemodynamic predictions. Understanding the influence of variation in image-derived properties is essential to assess the fidelity of model predictions. Numerous programs exist to render three-dimensional surfaces and construct vessel centerlines. Still, there is no exact way to generate vascular trees from the centerlines while accounting for uncertainty in data. This study introduces an innovative framework employing statistical change point analysis to generate labeled trees that encode vessel dimensions and their associated uncertainty from medical images. To test this framework, we explore the impact of uncertainty in 1D hemodynamic predictions in a systemic and pulmonary arterial network. Simulations explore hemodynamic variations resulting from changes in vessel dimensions and segmentation; the latter is achieved by analyzing multiple segmentations of the same images. Results demonstrate the importance of accurately defining vessel radii and lengths when generating high-fidelity patient-specific hemodynamics models.

Computational fluid dynamics modeling of coronary artery blood flow using OpenFOAM: Validation with the food and drug administration benchmark nozzle model

Cardiovascular disease (CVD), a global health concern, particularly coronary artery disease (CAD), poses a significant threat to well-being. Seeking safer and cost-effective diagnostic alternatives to invasive coronary angiography, noninvasive coronary computed tomography angiography (CCTA) gains prominence. This study employed OpenFOAM, an open-source Computational Fluid Dynamics (CFD) software, to analyze hemodynamic parameters in coronary arteries with serial stenoses. Patient-specific three-dimensional (3D) models from CCTA images offer insights into hemodynamic changes. OpenFOAM breaks away from traditional commercial software, validated against the FDA benchmark nozzle model for reliability. Applying this refined methodology to seventeen coronary arteries across nine patients, the study evaluates parameters like fractional flow reserve computed tomography simulation (FFRCTS), fluid velocity, and wall shear stress (WSS) over time. Findings include FFRCTS values exceeding 0.8 for grade 0 stenosis and falling below 0.5 for grade 5 stenosis. Central velocity remains nearly constant for grade 1 stenosis but increases 3.4-fold for grade 5 stenosis. This research innovates by utilizing OpenFOAM, departing from previous reliance on commercial software. Combining qualitative stenosis grading with quantitative FFRCTS and velocity measurements offers a more comprehensive assessment of coronary artery conditions. The study introduces 3D renderings of wall shear stress distribution across stenosis grades, providing an intuitive visualization of hemodynamic changes for valuable insights into coronary stenosis diagnosis.

A multi-stage neural network approach for coronary 3D reconstruction from uncalibrated X-ray angiography images

We present a multi-stage neural network approach for 3D reconstruction of coronary artery trees from uncalibrated 2D X-ray angiography images. This method uses several binarized images from different angles to reconstruct a 3D coronary tree without any knowledge of image acquisition parameters. The method consists of a single backbone network and separate stages for vessel centerline and radius reconstruction. The output is an analytical matrix representation of the coronary tree suitable for downstream applications such as hemodynamic modeling of local vessel narrowing (i.e., stenosis). The network was trained using a dataset of synthetic coronary trees from a vessel generator informed by both clinical image data and literature values on coronary anatomy. Our multi-stage network achieved sub-pixel accuracy in reconstructing vessel radius (RMSE = 0.16 ± 0.07 mm) and stenosis radius (MAE = 0.27 ± 0.18 mm), the most important feature used to inform diagnostic decisions. The network also led to 52% and 38% reduction in vessel centerline reconstruction errors compared to a single-stage network and projective geometry-based methods, respectively. Our method demonstrated robustness to overcome challenges such as vessel foreshortening or overlap in the input images. This work is an important step towards automated analysis of anatomic and functional disease severity in the coronary arteries.

Noninvasive computed tomography derived fractional flow reserve simulation based on microvascular tree model reconstruction

To establish a novel method for noninvasive computed tomography derived fractional flow reserve (CT-FFR) simulation based on microvascular tree model reconstruction and to evaluate the feasibility and diagnostic performance of the novel method in coronary artery disease compared with invasive fractional flow reserve (FFR). Twenty patients (20 vessels) who underwent coronary computed tomography angiography (CCTA) and invasive FFR were retrospectively studied. The anatomic epicardial coronary artery model was reconstructed based on CCTA image, and the microvascular tree model was simulated based on patient-specific anatomical structures and physiological principles. Numerical simulation was subsequently performed using the CFD method with full consideration of the variation of viscosity in microvascular. Two patients with the FFR value of .80 were selected for adjusting the parameters of the model, while the remaining 18 patients were selected as a validation cohort. After simulation, CT-FFR was compared with invasive FFR with a threshold of .80. Eleven (55%) patients had an abnormal FFR that was less than or equal to .80. Compared with invasive FFR, sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of CT-FFR with an optimal threshold of .80 were 100%, 77.8%, 81.8%, 100%, 88.89%, respectively. There were a good correlation and consistency between CT-FFR and invasive FFR. Time per patient of CT-FFR analysis was less than 15 min. CT-FFR based on microvascular tree model reconstruction is feasible with good diagnostic performance. It requires a short processing time with excellent accuracy. Large multicenter prospective studies are required for further demonstrating the diagnostic performance of this novel model in myocardium ischemia evaluation.

A novel method for calculating CTFFR based on the flow ratio between stenotic coronary and healthy coronary

BACKGROUND

Epicardial coronary stenosis may lead to myocardial ischaemia, and the resulting obstructive coronary artery disease is one of the leading causes of death. CT-derived fractional flow reserve (CT-FFR) has been clinically shown to be an effective method for the noninvasive assessment of coronary artery stenosis. However, this method has the problem that the measurement result is affected by the selected measurement position.

OBJECTIVES

This study was to obtain a novel flow-based approach to coronary CTFFR (CTQFFR), which was not affected by the measurement location.

METHODS

This study established healthy-assumed coronary arteries based on narrowed coronary arteries. Based on the assumption that the microvascular resistance remains unchanged in the short term after coronary stenosis treatment, the blood flow in the stenotic coronary artery and the healthy-assumed coronary artery was obtained by numerical simulation, and the CTQFFR based on the blood flow ratio was calculated. The functional relationship between CTQFFR and FFR was fitted by the results of 20 cases.

RESULTS

In this study, the functional relationship between CTQFFR and FFR was fitted by a quadratic curve, and the variance was 0.8744; the functional relationship between CTQFFR and pressure-based approach to coronary CTFFR (CTPFFR) was fitted by a primary curve, and the variance was 0.9971. There was coronary artery growth in all 20 cases. Preliminary validation results using 10 cases showed 100% accuracy in determining whether coronary artery stenosis required for clinical intervention. The relative error of the coefficient with the results proposed in a previous study was 0.316%.

CONCLUSION

This study proposes a new method for calculating coronary CTFFR, namely, coronary CTQFFR, which is the flow ratio between stenotic coronary and healthy-assumed coronary. This method solves the problem that the downstream CTFFR of coronary stenosis is related to the selected location, which effectively improves the CTFFR at the critical value (CTFFR= 0.8) near reliability. Preliminary research results show that the method obtained in this study has a high accuracy for determining whether there is significant coronary stenosis. However, large multi-centre validation for the feasibility of this method was necessary in our future work.

A predictive patient-specific computational model of coronary artery bypass grafts for potential use by cardiac surgeons to guide selection of graft configurations

Cardiac surgeons face a significant degree of uncertainty when deciding upon coronary artery bypass graft configurations for patients with coronary artery disease. This leads to significant variation in preferred configuration between different surgeons for a particular patient. Additionally, for the majority of cases, there is no consensus regarding the optimal grafting strategy. This situation results in the tendency for individual surgeons to opt for a “one size fits all” approach and use the same grafting configuration for the majority of their patients neglecting the patient-specific nature of the diseased coronary circulation. Quantitative metrics to assess the adequacy of coronary bypass graft flows have recently been advocated for routine intraoperative use by cardiac surgeons. In this work, a novel patient-specific 1D-0D computational model called “COMCAB” is developed to provide the predictive haemodynamic parameters of functional graft performance that can aid surgeons to avoid configurations with grafts that have poor flow and thus poor patency. This model has significant potential for future expanded applications.

The Continuum of Invasive Techniques for the Assessment of Intermediate Coronary Lesions

Ischemic heart disease is one of the most important causes of mortality and morbidity worldwide. Revascularization of coronary stenosis inducing ischemia, either by percutaneous or surgical intervention, significantly reduces major adverse cardiovascular events and improves quality of life. However, in cases of intermediate lesions, classified by a diameter stenosis between 50 and 90% by European guidelines and 40–70% in American counterparts with no clear evidence of ischemia, the indication of revascularization and impact is determined using various methods that altogether comprehensively evaluate the lesions. This review will discuss the various techniques to assess intermediate stenoses, highlighting indications and advantages, but also drawbacks. Fractional flow rate (FFR) and instantaneous wave-free ratio (iFR) are the gold standard for the functional evaluation of intermediate lesions, but there are clinical circumstances in which these pressure-wire-derived indices are not accurate. Complementary invasive investigations, mainly intravascular ultrasound and/or optical coherence tomography, offer parameters that can be correlated with FFR/iFR and additional insights into the morphology of the plaque guiding the eventual percutaneous intervention in terms of length and size of stents, thus improving the outcomes of the procedure. The development of artificial intelligence and machine learning with advanced algorithms of prediction will offer multiple scenarios for treatment, allowing real-time selection of the best strategy for revascularization.

A modified method of noninvasive computed tomography derived fractional flow reserve based on the microvascular growth space

OBJECTIVE

To establish a modified method for optimizing outlet boundary conditions (BC) of computed tomography-derived fractional flow reserve (CT-FFR), considering the myocardium as a growth space for microcirculation. The feasibility and diagnostic performance of the modified method in stable coronary artery disease (CAD) were compared with invasive fractional flow reserve (FFR).

METHODS

Nineteen patients (19 lesions) underwent coronary computed tomography angiography (CCTA) and following invasive FFR were included. The microcirculation resistance model generated based on patient-specific anatomical structures and physiological principles was used as the outlet BC, considering the myocardium as a growth space. Brachial artery pressure (BAP) plus or minus 10 mmHg was used as the inlet pressure BC to investigate the effect of the circadian rhythm. After simulation, CT-FFR was compared with invasive FFR with a threshold of 0.80.

RESULTS

Compared with invasive FFR, sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of CT-FFR with an optimal threshold of 0.80 were 100%, 100%, 100%, 100%, 100%, respectively. There were a good correlation and consistency between CT-FFR and invasive FFR. Little effect of the circadian fluctuation of BAP was found on the simulation.

CONCLUSIONS

A modified method for CT-FFR with high diagnostic accuracy compared with invasive FFR was established, considering the whole myocardial as the growth space for microcirculation. Circadian fluctuations in BAP could be ignored when it was used as the inlet BC.

Towards enabling a cardiovascular digital twin for human systemic circulation using inverse analysis

An exponential rise in patient data provides an excellent opportunity to improve the existing health care infrastructure. In the present work, a method to enable cardiovascular digital twin is proposed using inverse analysis. Conventionally, accurate analytical solutions for inverse analysis in linear problems have been proposed and used. However, these methods fail or are not efficient for nonlinear systems, such as blood flow in the cardiovascular system (systemic circulation) that involves high degree of nonlinearity. To address this, a methodology for inverse analysis using recurrent neural network for the cardiovascular system is proposed in this work, using a virtual patient database. Blood pressure waveforms in various vessels of the body are inversely calculated with the help of long short-term memory (LSTM) cells by inputting pressure waveforms from three non-invasively accessible blood vessels (carotid, femoral and brachial arteries). The inverse analysis system built this way is applied to the detection of abdominal aortic aneurysm (AAA) and its severity using neural networks.

High precision invasive FFR, low-cost invasive iFR, or non-invasive CFR?: optimum assessment of coronary artery stenosis based on the patient-specific computational models

The objective of this paper is to apply computational fluid dynamic (CFD) as a complementary tool for clinical tests to not only predict the present and future status of left coronary artery stenosis but also to evaluate some clinical hypotheses. In order to assess the present status of the coronary artery stenosis severity, and thereby selecting the most appropriate type of treatment for each patient, fractional flow reserve (FFR), instantaneous wave free-ratio (iFR), and coronary flow reserve (CFR) are calculated. To examine FFR, iFR, and CFR results, the effect of geometric features of stenoses, including diameter reduction (%), lesion length (LL), and minimum lumen diameter (MLD), is studied on them. It is observed that FFR is a more conservative index than iFR and CFR to assess the severity of coronary stenosis. In addition, it is seen that FFR, iFR, and CFR decrease by increasing LL and decreasing MLD. Therefore, the morphological indices, LL/MLD and LL/MLD̂4, with the calculated conservative cut-off values equal to 5.5 and 3.6, are considered. Next, some controversial clinical hypotheses about the assessment of the severity of coronary stenosis are evaluated numerically. These include the examination of FFR, iFR, and CFR accuracies, investigating the effect of coronary hyperemia on iFR, as well as the reliability of the hybrid iFR-FFR decision-making strategy. The presented numerical model can also be used as a predictive tool to identify the atherosuseptible sites of arteries by calculating the time-averaged wall shear stress (TAWSS), oscillatory shear index (OSI), and relative residence time (RRT).

Artificial intelligence approaches to predict coronary stenosis severity using non-invasive fractional flow reserve

Fractional flow reserve is the current reference standard in the assessment of the functional impact of a stenosis in coronary heart disease. In this study, three models of artificial intelligence of varying degrees of complexity were compared to fractional flow reserve measurements. The three models are the multivariate polynomial regression, which is a statistical method used primarily for correlation; the feed-forward neural network; and the long short-term memory, which is a type of recurrent neural network that is suited to modelling sequences. The models were initially trained using a virtual patient database that was generated from a validated one-dimensional physics-based model. The feed-forward neural network performed the best for all test cases considered, which were a single vessel case from a virtual patient database, a multi-vessel network from a virtual patient database, and 25 clinically invasive fractional flow reserve measurements from real patients. The feed-forward neural network model achieved around 99% diagnostic accuracy in both tests involving virtual patients, and a respectable 72% diagnostic accuracy when compared to the invasive fractional flow reserve measurements. The multivariate polynomial regression model performed well in the single vessel case, but struggled on network cases as the variation of input features was much larger. The long short-term memory performed well for the single vessel cases, but tended to have a bias towards a positive fractional flow reserve prediction for the virtual multi-vessel case, and for the patient cases. Overall, the feed-forward neural network shows promise in successfully predicting fractional flow reserve in real patients, and could be a viable option if trained using a large enough data set of real patients.

Comparison of Instantaneous Wave-Free Ratio (iFR) and Fractional Flow Reserve (FFR) with respect to Their Sensitivities to Cardiovascular Factors: A Computational Model-Based Study

While coronary revascularization strategies guided by instantaneous wave-free ratio (iFR) are, in general, noninferior to those guided by fractional flow reserve (FFR) with respect to the rate of major adverse cardiac events at one-year follow-up in patients with stable angina or an acute coronary syndrome, the overall accuracy of diagnosis with iFR in large patient cohorts is about 80% compared with the diagnosis with FFR. So far, it remains incompletely understood what factors contribute to the discordant diagnosis between iFR and FFR. In this study, a computational method was used to systemically investigate the respective effects of various cardiovascular factors on FFR and iFR. The results showed that deterioration in aortic valve disease (e.g., regurgitation or stenosis) led to a marked decrease in iFR and a mild increase in FFR given fixed severity of coronary artery stenosis and that increasing coronary microvascular resistance caused a considerable increase in both iFR and FFR, but the degree of increase in iFR was lower than that in FFR. These findings suggest that there is a high probability of discordant diagnosis between iFR and FFR in patients with severe aortic valve disease or coronary microcirculation dysfunction.